I sure hope this doesn't mean a move away from release formalism.

Github -- due in large part to complex social cues around encouraging public forking in a top-level namespace -- seems to encourage the assumption that developers will track git and add their own local patches, resulting in a significant decrease in upstream-interest in producing stable, formal, documented, tested, trustable releases and release builds.

Managing open source projects is difficult, and I too hope the quality and sensibility of releases doesn't degrade. But, Github might make the process more social, and easier to manage. More local patches could translate to more upstream patches.

> But, Github might make the process more social, and easier to manage. More local patches could translate to more upstream patches.

I hope it does. However, when Clojure switched to github, basically nothing changed. The only difference was that changes showed up immediately instead of having to wait for the cron job that ran git-svn. They still only accept patches as attachments to Jira issues.

So it totally depends on the maintainers attitude. If they are switching because they're tired of everyone bugging them about it, then you probably won't see much change. If they're switching because they see compelling benefits to a process that allows more community involvement, then you might be in luck.

Perhaps, but this is much less of a deterrent for me than having to deal with their very slow svn server.

otoh, this also leads to more peer review, which in turn could contribute to all of parents points.

Now it will be easier to keep updated to the source. I can't wait to look inside the source on the nightly builds just to see the what and why of things first hand.

Finally!